Translates a text column in a Flink SQL job using a large language model (LLM).
Limitations
-
Requires Ververica Runtime (VVR) 11.4 or later.
-
To use the Flink AI service (built-in models), VVR 11.7 or later is required, and the Flink AI service must be activated. For details, see Flink AI service (built-in models).
-
Throughput is subject to rate limiting by the model platform. When the rate limit is exceeded, the Flink job experiences backpressure and this operator becomes a bottleneck. Severe rate limiting triggers timeout errors and causes the job to restart.
Syntax
AI_TRANSLATE(
MODEL => MODEL <MODEL NAME>,
INPUT => <INPUT COLUMN NAME>,
SOURCE_LANG => <SOURCE LANGUAGE>,
TARGET_LANG => <TARGET LANGUAGE>
)Parameters
| Parameter | Data type | Description |
|---|---|---|
MODEL <MODEL NAME> |
MODEL | The name of the registered model service. For more information, see Model settings. The output type of this model must be VARIANT. |
<INPUT COLUMN NAME> |
STRING | The source text to translate. |
<SOURCE LANGUAGE> |
STRING | The language code of the source text (for example, zh for Chinese). Pass auto to let the model detect the language automatically. Must be a constant. |
<TARGET LANGUAGE> |
STRING | The language code of the target language (for example, zh for Chinese). Must be a constant. |
The supported languages are determined by the model you register.
Supported languages
| Language | Code |
|---|---|
| Arabic | ar |
| Chinese | zh |
| English | en |
| French | fr |
| German | de |
| Japanese | ja |
| Korean | ko |
| Portuguese | pt |
| Russian | ru |
| Spanish | es |
Return values
| Field | Data type | Description |
|---|---|---|
translated_text |
STRING | The translated text. |
detected_language |
STRING | The detected source language. |
Examples
The following examples reference a Flink built-in model and use AI_TRANSLATE to translate English text to Chinese. Both positional and named argument syntax are shown.
Test data
| id | content |
|---|---|
| 1 | I like Flink very much. It's so fast! |
SQL statements
CREATE TEMPORARY MODEL general_model
INPUT (`input` STRING)
OUTPUT (`content` VARIANT)
WITH (
'provider' = 'openai-compat',
'task' = 'chat/completions',
'model' = 'qwen3.6-flash'
);
CREATE TEMPORARY VIEW infos(id, content)
AS VALUES (1, 'I like Flink very much. It''s so fast!');
-- Positional argument syntax
SELECT id, translated_text, detected_language
FROM infos,
LATERAL TABLE(
AI_TRANSLATE(
MODEL general_model,
content,
'auto',
'zh'
));
-- Named argument syntax
SELECT id, translated_text, detected_language
FROM infos,
LATERAL TABLE(
AI_TRANSLATE(
MODEL => MODEL general_model,
INPUT => content,
SOURCE_LANG => 'auto',
TARGET_LANG => 'zh'
));Output
| id | translated_text | detected_language |
|---|---|---|
| 1 | I like Flink very much. It's very fast! | en |